In partnership with
Welcome to the iPrompt Newsletter
Friday Edition
What you get in the Friday Edition
Weekly Scoreboard
Top Headlines of the Week
Our Investing Angle
Three Ideas to Research This Weekend
AI Investment Framework
Your Move
…all in a FREE Weekly newsletter.
Sponsor:
Investors see ANOTHER return from Masterworks (!!!!)
That’s 6 sales in 7 months. 29 all time. And the performance?
16.5%, 17.6%, and 17.8%, net annualized returns on sold works held longer than one year (See all 29 at Masterworks.com)
It’s not from stocks, private equity, or real estate… it’s from contemporary and post war art. Crazy, right?
With Masterworks, you don’t need to be a BILLIONAIRE to invest in multi-million dollar art anymore.
Historically, the segment overall has had attractive appreciation and low correlation to stocks.*
Masterworks targets works featuring legends like Banksy, Basquiat, and Picasso, identifying what they believe to have significant long-term appreciation potential, not just at the artist level but at the level of individual artworks.
As one of the largest players in the art market, with $1.3 billion invested over 500 artworks, they pass critical advantages through to their 70,000+ members to add art to their portfolios strategically.
Looking to diversify your investments in 2026?
*According to Masterworks data. Investing involves risk. Past performance is not indicative of future returns. See important Reg A disclosures at masterworks.com/cd.
iPrompt Signals
AI & robotics investing — explained so you can actually act on it.
ISSUE 12 // Friday, 29 May 2026 // 6 min read
The Hook
Last week the whole sector hung on one number: the 30-year Treasury yield at a 19-year high. This week it eased — and here's the surprise. The relief money didn't pile back into Nvidia. It rotated one shelf down the chip stack, into memory. When the rate fear lifts, you find out what the market actually wanted to own. This week the answer wasn't the chip everyone watches — it was the chip stacked next to it.
🎯 YOUR ONE DECISION THIS WEEKEND Three lenses on the rotation. Pick one to research — not to buy. 1. Memory (MU) — is the HBM re-rating a structural shift, or the last, hottest leg of the cycle? 2. Foundry (TSM) — Nvidia's $150bn-a-year Taiwan spend has to land somewhere. Who banks it? 3. Duration unwind — if yields keep falling, do last week's avoided high-multiple names lead the bounce? The deliverable: one ticker, one tripwire, before Monday's open. The three research paths below do the heavy lifting. |
Weekly Scoreboard
Ticker | Price (28 May) | Week % | What Happened |
MU | $942.36 | +23.6% | Memory melt-up. UBS tripled its target. |
TSM | $424.86 | +4.3% | Nvidia's $150bn-a-year Taiwan pledge. |
AVGO | $426.58 | +3.5% | Custom silicon; Q2 print 5 June. |
NVDA | $212.58 | −3.2% | Perfect quarter, still drifting. |
BOTZ | $40.7 | +1.5% | Robotics steady; physical-AI read-through. |
S&P 500 | 7,520 | +1.2% | New record; eased rates did the work. |
US 30Y | 5.03% | −16 bp | Pulled back from a 19-year high. |
VIX | 16.0 | calm | Fear gauge back to summer-lull levels. |
Bottom line: the bond market eased off the throat — and the money rotated one shelf down the chip stack, into memory, rather than back into the old leader. |
VIX = the “fear gauge”: expected S&P 500 swings over the next 30 days. Above 25 = real anxiety; 16 = relaxed.
The Two Stories That Matter
1. Three memory makers, one trillion-dollar week.
The hard facts first. On Tuesday, UBS roughly tripled its Micron price target to a street-high $1,625; the stock jumped about 19% — its biggest day since 2011 — and crossed $1tn in market value (UBS note, reported by CNBC and Bloomberg, 26 May). On Wednesday SK Hynix did the same (Reuters, 27 May), joining Samsung, which crossed earlier in May. Three of the world's memory makers, a trillion dollars each, inside one month. What it means: the market is re-rating memory from a boom-bust commodity to a structural AI bottleneck — a different multiple, not just a better quarter.
🌱 New to investing? (Know HBM already? Skip this box.) Every AI chip needs memory chips beside it to feed data fast enough — the specialist kind is HBM (high-bandwidth memory). Memory used to be a cheap commodity: prices boomed and crashed, so the stocks traded on a low multiple. The bet this week is that AI demand broke that cycle — supply is so tight that prices stay high for years. If true, the same earnings deserve a higher price. That re-rating, not one good quarter, is what a trillion-dollar valuation prices in. |
2. Nvidia drifts while it pledges $150bn a year to Taiwan.
The fact: at the groundbreaking for “Constellation,” its first overseas headquarters in Taipei, Jensen Huang said Nvidia now spends about $100bn a year in Taiwan and is heading for $150bn — up roughly tenfold in four or five years (Huang, speaking at the launch event; Reuters and CNBC, 27 May). Taiwan's index closed at a record; Nvidia's own shares drifted about 3% lower on the week. My read: the week's biggest capital commitment didn't lift the company that made it — it lifted the supply chain it flows into. The value is leaking downstream, to the foundry that builds the chips and the memory that goes beside them.
Supporting signals
(context for sizing, not main trades)
SpaceX files to go public (S-1, ~$1.75–2tn target, SPCX, ~12 June window; Fortune/Bloomberg, 20–28 May): the purest “2030 promise” on the board — orbital AI compute, big losses — about to ask public markets to fund it. The cleanest test of last week's duration thesis.
ByteDance's ~$70bn data-centre push (TipRanks/CNBC, 27 May): more confirmation the demand under the memory trade is real — and global.
Broadcom reports Q2 on 5 June (calendar): the custom-silicon read on whether hyperscaler design-wins keep compounding. Watch the AI-revenue guide, not the headline beat.
Our Investing Angle
The stories are above. This section just bridges them to the research paths: who benefits, who's exposed, what to watch.
What changed: the AI trade rotated down the stack, from compute to memory. The most-owned name printed a perfect quarter and a $150bn pledge and still fell, while the supply-constrained part of the build-out went vertical. When the rate fear lifted, capital went looking for genuine scarcity — and, if the re-rating holds, that scarcity is in high-bandwidth memory.
Who benefits: the bottleneck owners — Micron, SK Hynix, Samsung — plus TSM, the foundry every chip passes through and the banker of Nvidia's Taiwan spend. Who's exposed: the high-multiple software names that need rates to keep falling, Palantir foremost among them. They have no special cover if yields spike again. What to watch: two lines decide whether this rotation holds — memory contract pricing, and the 30-year yield.
→ Deep dive: The AI memory map — why HBM became the chokepoint, how the three makers split the pie, and the signals that separate a re-rating from a cycle top. Read it →
⚠️ WHAT COULD GO WRONG? (the bear case) 1. Memory is still a cycle. Three names crossed $1tn in a month and Micron is up roughly 225% year-to-date (price data, 28 May). Even if AI breaks the boom-bust, buying the vertical part of the chart is how cycles end. → Argues for waiting for a pricing data point before sizing up. 2. The rate reprieve may be temporary. A BofA survey had 62% of managers expecting the 30-year to hit 6%. One hot inflation print and last week's scare is back. → Argues for the foundry (self-funding) over the highest-multiple names. 3. Nvidia drifting isn't always bullish for the stack. If the leader fades because demand is finally being questioned — not just rotation — then memory and foundry are late-cycle longs. → Argues for smaller size until Broadcom's 5 June print clarifies demand. |
Your Three Research Paths
The core of this issue. Not recommendations — starting points. Pick one. Tighten it before Monday.
Path 1 — Memory (MU). The bottleneck the whole stack now depends on.
Why now: Micron jumped ~19% on a UBS street-high $1,625 target and crossed $1tn the same week SK Hynix and Samsung did — the market re-rating memory from commodity to AI chokepoint in real time.
The case: HBM sits beside every AI GPU, and Micron's own management has called supply structurally tight through 2027 (its last earnings call). Tight supply plus locked-in pricing is margin — and margin is what a re-rating is built on.
The risk: Memory has always been cyclical, and this is the vertical part of the chart. A re-rating that's already happened isn't an entry — it's a position you're chasing.
Tripwire: DRAM/HBM contract pricing rolling over at Micron's 24 June print = the cycle still rules and the re-rating overshot. Pricing holding or rising into 2027 guidance = the structural-bottleneck thesis is live.
How to research: MU directly, or the Roundhill memory ETF for the basket. Read the pricing commentary in the call, not the headline EPS.
Path 2 — Taiwan Semiconductor (TSM). The one who banks Nvidia's $150bn.
I keep coming back to TSM because it's the cleanest way to own the rotation without picking which shelf wins. Why now: Nvidia just said it's heading for $150bn a year in Taiwan, and TSM is the foundry that fabricates essentially every leading-edge AI chip, Nvidia's included. It funds its own capex from cash flow, so it barely cares where the 30-year sits.
The case: Whether the money rotates to compute, memory, or custom silicon, it passes through TSM's fabs. And it trades at a discount to most US-listed AI names — roughly 26x forward earnings (Bernstein/247WallSt, late May) — despite owning the chokepoint.
The risk: Taiwan geopolitical risk is permanent and binary — and Nvidia's deepening concentration on the island makes the whole complex more exposed, not less.
Tripwire: TSM raising its 2026 capex guide again at the 16 July print = demand is outrunning even the upgraded plan. Any formal Section 232 chip-tariff timeline from Washington = re-price the whole Taiwan complex before acting.
How to research: TSM (US ADR) or 2330.TW. Compare against SOXX/SMH for the diversified version.
Path 3 — The duration unwind. The lens, not the trade.
Look, I'm not fully sold on this one — but it's the cleanest test of last week's whole thesis. Why now: the 30-year fell ~16bp and the high-multiple names I told you to avoid didn't obviously lead the bounce — memory did. So either the duration trade is unwinding quietly, or it isn't unwinding at all.
The case: If yields keep easing, the longest-duration AI assets — Palantir, the application-software layer, even a 2030 promise like SpaceX — are mechanically the biggest winners. This is a lens for sizing your other positions, not a single trade.
The risk: One week of falling yields is noise, not a trend. 62% of fund managers still expect 6% on the long bond. Lean into this and a hot print takes it straight back.
Tripwire: 30-year closing back below 4.75% before the 18 June Fed meeting = the unwind is real and the high-multiple names are the leveraged read. Back above 5.25% = the scare isn't over, and the case for the foundry/memory lenses is the one worth re-examining.
How to research: Watch TLT (long-Treasury ETF) as the live read on yields; PLTR as the purest high-duration AI single name.
AI Investment Framework
Living portfolio framework — the full layer-by-layer breakdown lives on the site. Here's the dashboard, the rotation in one chart, and what changed this week. Not financial advice; research starting points only.
Six months in one line: the layers held through April, the long-bond scare clipped everything in late May, and the snapback came led by memory and the foundry — not the old single-name leader.
Changes this week
Infrastructure held HIGH ↔ — but the leadership inside it moved from Nvidia to memory and foundry. Same signal, new leader.
Global raised DEVELOPING ↓ → DEVELOPING ↑↑. SK Hynix crossing $1tn and a record KOSPI make the AI story genuinely East-Asian this week, not just American.
Applications nudged MEDIUM ↑ on easing yields — a tailwind from rates, not fundamentals. Risk stays at 5/5.
Two dates that matter
Date | Event | What It Resolves |
Thu 5 Jun | Broadcom Q2 earnings | AI-revenue guide and custom-silicon design-wins — the read on whether the build-out keeps compounding. |
Fri 12 Jun | SpaceX IPO (SPCX) window | The first public vote on “AI infrastructure in space” — and on funding 2030 promises at a 2026 yield. |
Disclaimer: This newsletter is for informational and educational purposes only and does not constitute financial advice. iPrompt Signals is not a registered investment advisor. Always conduct your own research and consult a qualified financial professional before making investment decisions.
Your Move
Pick one of the three research paths. 45 minutes, latest filing, your own tripwire. Then hit reply with the one you're taking — memory, the foundry that banks it, or the rate unwind underneath it all.
🌱 SHORT TAKE For the broad-exposure reader: this week was rotation, not relief. If you don't want to call which shelf of the chip stack leads next, SMH (VanEck Semiconductors) holds Nvidia, the foundry, and the memory names together — you own the rotation instead of guessing it. Not a recommendation — a starting point. |
Stay curious — and stay qualified.
— R. Lauritsen
Editor, iPrompt Signals
Know someone building an AI position? Forward this — they'll thank you by Friday.
P.S. — Last Friday I called the duration trade right: yields did reverse. But I had the winner wrong. I expected the high-multiple names to lead the bounce; instead the relief went straight into memory. Being right about the move and wrong about who'd win from it is the part worth sitting with.
→ Deep dive: The AI memory map — why HBM became the chokepoint, how the three makers split the pie, and the signals that separate a re-rating from a cycle top. Read it →
Terms Used This Week
HBM (high-bandwidth memory) — The specialist memory stacked beside an AI GPU to feed it data fast enough. The bottleneck behind this week's memory melt-up.
Re-rating — When the market decides a business deserves a higher (or lower) multiple on the same earnings. Memory going from commodity multiple to AI-chokepoint multiple is a re-rating.
Foundry — A factory that makes chips designed by other companies. TSM is the foundry behind essentially every leading-edge AI chip.
Duration — How far in the future an asset's cash flows sit. Long-duration assets (high-multiple software, a 2030 promise) move most when rates change.
Section 232 — A US trade law that lets the government impose tariffs on national-security grounds — here, the lever that could tax imported chips.
Sponsor:
The LA Mayor Odds Are Moving. June 2 Is the Deadline.
Bass at 68%, Pratt at 27% — and $21M already trading on the outcome. The best trades happen before the picture gets clear. Get in now and get $10 free to start.
Trade responsibly.